Table 1 k-mer embedding encodes more accurate motif information.
Feature Extractor | k-mer | AUC (%) | ACC (%) | MCC (%) | Sensitivity (%) | Specificity (%) | Precision (%) | F1 (%) |
|---|---|---|---|---|---|---|---|---|
One hot encoding | 1 | 82.31 | 78.19 | 56.59 | 81.18 | 75.41 | 75.41 | 78.19 |
2 | 87.07 | 80.74 | 61.41 | 79.41 | 81.97 | 80.36 | 79.88 | |
3 | 87.66 | 83.29 | 66.66 | 78.24 | 87.98 | 85.81 | 81.85 | |
4 | 86.00 | 79.60 | 59.39 | 82.35 | 77.05 | 76.92 | 79.55 | |
5 | 85.57 | 80.17 | 60.27 | 78.82 | 81.42 | 79.76 | 79.29 | |
6 | 87.52 | 80.17 | 60.27 | 77.65 | 82.51 | 80.49 | 79.04 | |
Word2Vec | 1 | 86.91 | 78.75 | 59.99 | 91.76 | 66.67 | 71.89 | 80.62 |
2 | 83.38 | 76.20 | 53.78 | 61.76 | 89.62 | 84.68 | 71.43 | |
3 | 87.17 | 80.17 | 60.27 | 78.24 | 81.94 | 80.12 | 79.17 | |
4 | 68.50 | 64.02 | 29.35 | 74.71 | 54.10 | 60.19 | 66.67 | |
5 | 62.41 | 56.37 | 12.21 | 41.18 | 70.49 | 56.45 | 47.62 | |
6 | 60.05 | 58.36 | 16.87 | 60.59 | 56.28 | 56.28 | 58.36 | |
GloVe | 1 | 85.12 | 78.19 | 56.38 | 78.82 | 77.60 | 76.57 | 77.68 |
2 | 84.35 | 76.49 | 53.41 | 81.18 | 72.13 | 73.02 | 76.88 | |
3 | 84.47 | 80.17 | 60.83 | 71.18 | 88.52 | 85.21 | 77.56 | |
4 | 59.21 | 57.79 | 18.45 | 21.18 | 91.80 | 70.59 | 32.58 | |
5 | 61.62 | 60.06 | 19.90 | 56.47 | 63.39 | 58.90 | 57.66 | |
6 | 62.73 | 59.77 | 19.38 | 57.06 | 62.30 | 58.43 | 57.06 | |
fastText | 1 | 84.71 | 76.77 | 53.75 | 68.82 | 84.15 | 80.14 | 74.05 |
2 | 83.64 | 75.35 | 53.40 | 89.41 | 62.30 | 68.78 | 89.41 | |
3 | 85.36 | 77.05 | 55.32 | 85.88 | 68.85 | 71.92 | 78.28 | |
4 | 64.29 | 61.76 | 24.02 | 41.76 | 80.33 | 66.36 | 51.26 | |
5 | 61.52 | 57.22 | 16.40 | 73.53 | 42.08 | 54.11 | 62.54 | |
6 | 61.40 | 55.81 | 11.04 | 37.06 | 73.22 | 56.25 | 44.68 |
